Vladislav Zaimov brings a wealth of knowledge to the table regarding the evolving landscape of enterprise telecommunications. As an expert in managing risks within vulnerable networks, he understands the delicate balance between adopting cutting-edge technology and maintaining the ironclad reliability that users expect. Today, we explore how agentic AI is transforming the industry from a cost-saving measure into a powerful engine for revenue growth and operational excellence. This discussion moves beyond simple automation to examine how carriers can leverage their physical proximity to users and their vast network data to establish a new “North Star” for the digital era.
Telecom operators possess unique physical assets like power and network sites that sit close to end users, but how does this specific geographic advantage translate into a win for agentic AI implementation?
The proximity of telecom infrastructure to the actual consumer is perhaps the most significant advantage operators have as AI shifts away from centralized data centers. Because these providers already own the power sources, physical sites, and established customer relationships, they are uniquely positioned to monetize AI at the edge where low latency is critical. By deploying agentic AI closer to the user, carriers can offer faster response times and more personalized services that third-party tech giants simply cannot match without that same ground-level footprint. This physical foundation allows for scaling networks in an autonomous manner, ensuring that services are delivered with a speed that matches the real-time demands of modern enterprise applications.
Many organizations struggle with scattered pilot programs that never seem to scale, so what should a “North Star architecture” look like for a carrier trying to build a truly autonomous network?
A successful North Star architecture requires moving away from disconnected experiments and toward a unified platform model for network operations. At events like DTW Ignite in Copenhagen, the consensus has been that operators must build a foundation of operational efficiency that allows them to go to market faster rather than just chasing small cost reductions. This architectural vision acts as a guiding light, ensuring that every AI agent deployed contributes to a cohesive system rather than creating another silo of technology. By establishing this clear target architecture early on, teams gain the room they need to prove value quickly through early use cases while still working toward a fully autonomous, scalable ecosystem.
Data is often cited as the biggest hurdle for legacy carriers, with information trapped in old systems, so how can agents help bridge the gap without waiting for a perfect data migration?
Waiting for perfect data is a trap that many carriers fall into, but the current strategy is to meet the data exactly where it resides today. Data retrieval agents are becoming a practical starting point because they can reach into separate, aging platforms to pull the necessary insights for real-time decision-making. Instead of a massive, multi-year overhaul, these agents act as the connective tissue between legacy systems and modern AI frameworks. This approach allows operators to begin realizing the benefits of agentic AI immediately, using existing network data to find risks and diagnose issues without the need for a total data center transformation.
Given that a single error in a telecom network can disrupt critical services for thousands, how do multi-agent frameworks like Crosswork AI manage the risks of autonomous decision-making?
Multi-agent frameworks are designed to distribute operational tasks among specialized agents, which helps in isolating functions and reducing the blast radius of any potential error. For instance, one agent might be dedicated to finding risks, while another focuses specifically on diagnosing issues and a third suggests fixes, creating a system of checks and balances. This modularity is essential for maintaining high responsibility over enterprise and critical services where downtime is not an option. By using these frameworks, engineering teams can reduce the immense pressure they face, as the AI handles the repetitive diagnostic work while leaving the final high-stakes authorizations to human oversight.
Since tools must fit into real-world workflows to be effective, what are the primary challenges in change management when introducing AI agents as “team extensions”?
The human element is often a bigger hurdle than the technology itself, as engineers and operations leaders need to feel a high level of trust in the agents they manage. To build this trust, AI agents are increasingly being given their own identities, complete with specific guardrails and sets of policies that govern their actions. It is vital that these agents act according to established permissions, allowing human operators to understand exactly what an agent is doing and why at any given moment. When agents are treated as extensions of the team rather than black-box replacements, the integration into existing workflows becomes much smoother and more disciplined.
What is your forecast for the adoption of agentic AI in the telecom sector over the next few years?
I believe we are going to see a rapid shift where the winners in the industry are those who pair high ambition with extreme operational discipline. We will move past the era of disconnected pilots and see the rise of standard platforms where autonomous networks are governed as carefully as they are deployed. As carriers refine their “North Star” architectures, agentic operations will become the standard foundation for creating new services at a pace we haven’t seen before. Ultimately, the successful integration of these technologies will rely on robust governance and the ability to turn fragmented legacy data into a reliable stream of actionable intelligence.
